1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

The expression of the auxin responsive reporter construct, GH3:gusA, was examined in transgenic white clover plants to assess changes in the auxin balance during the earliest stages of root nodule formation. Reporter gene expression was monitored at marked locations after the application of bacteria or signal molecules using two precise inoculation techniques: spot-inoculation and a novel method for ballistic microtargeting. Changes in GH3:gusA expression were monitored after the inoculation of Rhizobium leguminosarum biovar trifolii, non-host rhizobia, lipo-chitin oligosaccharides (LCOs), chitin oligosaccharides, a synthetic auxin transport inhibitor (naphthylphthalamic acid; NPA), auxin, the ENOD40–1 peptide or different flavonoids. The results show that clover-nodulating rhizobia induce a rapid, transient and local downregulation of GH3:gusA expression during nodule initiation followed by an upregulation of reporter gene expression at the site of nodule initiation. Microtargeting of auxin caused a local and acropetal upregulation of GH3:gusA expression, whereas NPA caused local and acropetal downregulation of expression. Both spot-inoculation and microtargeting of R. l. bv. trifolii LCOs or flavonoid aglycones induced similar changes to GH3:gusA expression as NPA. O-acetylated chitin oligosaccharides caused similar changes to GH3:gusA expression as R. l. bv. trifolii spot-inoculation, but only after delivery by microtargeting. Non-O-acetylated chitin oligosaccharides, flavonoid glucosides or the ENOD40–1 peptide failed to induce any detectable changes in GH3:gusA expression. GH3:gusA expression patterns during the later stages of nodule and lateral root development were similar. These results support the hypothesis that LCOs and chitin oligosaccharides act by perturbing the auxin flow in the root during the earliest stages of nodule formation, and that endogenous flavonoids could mediate this response.


  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Legumes can engage in a symbiosis with soil bacteria of the genera Rhizobium, Azorhizobium, Sinorhizobium and Bradyrhizobium, resulting in the formation of nitrogen-fixing root nodules. These ‘rhizobia’ trigger the nodule developmental programme via the secretion of mitogenic lipo-chitin oligosaccharides, called LCOs, as reviewed by Dénariéet al. (1996), Schultze et al. (1994) and Spaink (1996). In plant roots, LCOs also trigger hair deformation, preinfection thread formation, cortical cell division, flavonoid induction and secretion, and induction of nodulin and cell cycle genes.

 In temperate legumes, such as clover and alfalfa, nodules originate from inner cortical cell divisions that mostly occur opposite protoxylem poles (Hirsch 1992). All cortical cells between a particular protoxylem pole and the epidermis are triggered into division, but only the inner cortical cells complete division, eventually forming a nodule primordium (Yang et al. 1994). Although externally applied LCOs trigger cortical cell divisions, the nature of the endogenous plant signals induced by LCOs prior to the first cell divisions remain unknown.

 Several studies suggest that LCOs work by affecting phytohormone regulation (Hirsch 1992). Early work suggested that transverse gradients of signal molecules from the root stele and from infecting bacteria overlap to induce cortical cell divisions (Libbenga et al. 1973). This model proposed that the nodule progenitor cells of the inner cortex are initially exposed to an optimal ratio and concentration of signal molecules before forming a nodule primordium. Nodules can also be induced by manipulating plant hormone levels (Bauer et al. 1996;Cooper and Long 1994) or by the addition of synthetic auxin transport inhibitors (Hirsch et al. 1989). Flavonoids, which can act as endogenous auxin transport regulators (Jacobs and Rubery 1988), are possible candidates to mediate the effect of LCOs on the plant hormone balance (Hirsch 1992;Hirsch et al. 1989). Recent studies suggest that the ENOD40–1 peptide, which is induced from a very early stage in nodule development, affects the response to auxin in tobacco protoplasts during division (van de Sande et al. 1996). The phytohormones cytokinins and auxins are most likely to be involved in nodule initiation, since they are necessary for activation and completion of the cell cycle (John et al. 1993;Zhang et al. 1996).

 To investigate the role of endogenous signal molecules in root nodulation, we monitored the inferred changes in the auxin balance occurring during nodulation in white clover using transgenic plants carrying the promoter of the auxin responsive gene, GH3, fused to gusA (Larkin et al. 1996). GH3 is a very early auxin responsive gene from soybean (Hagen et al. 1984;Napier and Venis 1995), and the reporter gene, gusA encodes β-glucuronidase (GUS) (Jefferson et al. 1987). Our earlier work indicated that endogenous GH3:gusA expression occurred in patterns that were consistent with the known role of auxin (Larkin et al. 1996), thereby constituting a useful marker to study hormone-mediated differentiation processes. For lateral root formation, the requirement for auxin has already been established (Sussex et al. 1995;Wightman et al. 1980). Therefore, we used lateral root formation as a control to examine the changes in auxin responsiveness during nodule formation in clover. We also compared the effect of LCOs and LCO derivatives upon GH3:gusA expression to that of compounds known to interfere with the plant auxin balance such as the auxin transport inhibitor, NPA. Two micro-inoculation techniques, spot-inoculation (Bhuvaneswari et al. 1981) and microtargeting (Sautter et al. 1991;Schlaman et al. 1997), were used to localize a microbe or compound to marked sites on, or inside, the root, respectively. Both these methods made it possible to relate any temporal or spatial changes in reporter gene expression to the inoculation site as a reference point. Our results suggest that nodulating rhizobia, clover-specific LCOs, NPA and flavonoid aglycones transiently inhibit auxin transport and that this leads to a transient accumulation of auxin at the site where a nodule will be initiated.


  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Localization of GH3:gusA in untreated roots detected by histochemical GUS staining and immunohistochemistry

A comparison of histochemical GUS staining and immunohistochemistry showed that histochemical GUS staining occurred only in the cells in which the GUS enzyme was located (Fig. 1a,b). In control roots, GH3:gusA expression and the GUS enzyme were localized in the vascular bundle with little or no expression in the cortex (Figs 1c and 2a). Expression inside the vascular bundle occurred in the pericycle and in the parenchyma, especially around the xylem poles (Fig. 2a).


Figure 1 Localization of GH3:gusA activity in roots of uninoculated and inoculated transgenic white clover plants. (a) and (b) show the distribution of GUS in subsequent longitudinal sections of an untreated root, located in the vascular bundle (between arrowheads). In (a) the distribution of GUS activity was detected by histochemical staining, whereas in (b) the GUS protein was detected with fluorescently (FITC) labelled anti-GUS antibodies (green). Non-specific yellow autofluorescence is seen in xylem elements. (c–i) Whole roots showing the effect of spot-inoculated R. l. bv. trifolii (ANU843) upon GH3:gusA expression during the first 50 h (stages 1–5) of nodule development. The inoculation site is marked with an ion exchange bead (arrowhead). (c) GH3:gusA expression was seen in the vascular bundle and, at lower levels, also in cortex cells in an untreated root. (d) One hour p.i. no changes in GH3:gusA expression were observed. (e) At stage 1 (5 h p.i.), strong reduction of GH3:gusA expression occurred in the cortex and vascular bundle acropetal from the inoculation site. (f) At stage 2 (10 h p.i.), the reduction of GH3:gusA expression acropetal from the inoculation site persisted, and an upregulation of GH3:gusA expression occurred basipetal from the inoculation site. (g) At stage 3 (20 h p.i.), GH3:gusA expression was enhanced in the cortex in a localized zone around the inoculation site and expression reappeared in the vascular bundle acropetal from the inoculation site. (h) At stage 4 (30 h p.i.), an extended zone of high GH3:gusA expression occurred around the inoculation site (transverse section shown in Fig. 2e). Compared to 20 h p.i., the zone of enhanced GH3:gusA expression was located further basipetal and stretched further acropetal from the inoculation site. In about 50% of cases, GH3:gusA expression just behind the root tip was still reduced. (i) At stage 5 (50 h p.i.), a nodule primordium developed at the site of inoculation. A transverse section of this primordium is shown in Fig. 2(f). The dividing cells expressed GH3:gusA, whereas the expression in flanking regions was not affected. Bars = 50 μm in (a) and (b); 2 mm in (c), (d), (e), and (f); 3 mm in (g) and (h); 1. mm in (i).

Download figure to PowerPoint


Figure 2 Expression of GH3:gusA during parallel stages of lateral root and nodule development. (a–d) Stages of lateral root development. p = pericycle; e = endodermis; c = cortex (a) In the root zone in which lateral root initiation had not yet started, GH3:gusA expression was localized in the vascular bundle, including undivided pericycle cells and parenchyma around the xylem poles (arrows). (b) The first dividing pericycle cells showed GH3:gusA expression, but no GH3:gusA expression was detected in the endodermis or the cortex ahead of the dividing cells. (c) After several rounds of cell divisions, the growing primordium showed no GH3:gusA expression. Dividing cells in the xylem parenchyma of the respective xylem pole (arrowheads) and cells in the cortex overlying the growing primordium expressed GH3:gusA. (d) In the emerged lateral root, cells of the newly formed central vascular cylinder expressed GH3:gusA. Some expression was present in the cortex of the main root at the site of breakthrough of the lateral root (arrowheads). (e–h) Stages of nodule development after spot-inoculation with ANU843, shown as transverse sections at the site of inoculation. (e) At stage 4 (30 h p.i.), enhanced GH3:gusA expression occurred in all cortical and vascular cells at the position of inoculation prior to cell divisions in the inner cortex. (f) At stage 5 (50 h p.i.), the first dividing inner cortex cells strongly expressed GH3:gusA, whereas adjacent cortex cells only showed very weak GH3:gusA expression. (g) At stage 6 (70–90 h p.i.), the growing nodule primordium showed GH3:gusA expression occurring at the side and base, but not in the centre of the primordium. (h) At stage 7 (4–5 days p.i.), GH3:gusA expression occurred in a bifurcated pattern (arrows) with no detectable GH3:gusA expression occurring in the centre of the young primordium. Bars = 50 μm in (a) and (c); 20 μm in (b); 80 μm in (d); 100 μm in (h); 12. 0 μm in (e); 40 μm in (f) and (g).

Download figure to PowerPoint

Expression of GH3:gusA during lateral root organogenesis

Undivided pericycle cells in the young root stele expressed GH3:gusA (Fig. 2a). The earliest dividing pericycle cells, which are likely to be the progenitors of a lateral root, maintained GH3:gusA expression (Fig. 2b). Expression of GH3:gusA in the dividing pericycle cells clearly preceded that seen in the cortex in front of lateral root primordia at a later stage of development. After several rounds of pericycle divisions, the expression of GH3:gusA in the lateral root primordium diminished and cortex cells overlying the lateral root primordium strongly expressed GH3:gusA (Fig. 2c). After lateral root emergence, the newly formed vascular bundle expressed GH3:gusA (Fig. 2d), thus re-establishing in the lateral root the same pattern of GH3:gusA expression seen in the main root. The cortex region, through which the lateral root had penetrated, retained GH3:gusA expression (Fig. 2d).

GH3:gusA expression during early nodule initiation and development

Roots spot-inoculated with strain ANU843, a wild-type Rhizobium leguminosarum biovar trifolii, showed temporal and spatial changes in GH3:gusA expression during different stages of nodule formation. The confined spot-inoculation of strain ANU843 resulted in a high (84%) probability of a nodule occurring at the inoculated site, which enabled us to confidently relate changes in GH3:gusA expression patterns to the inoculation point.

 The changes in GH3:gusA expression observed could be categorized into distinct stages. Compared to control roots (Fig. 1c), no changes in GH3:gusA expression occurred within 1 h of inoculation (Fig. 1d). Between 1 and 5 h post-inoculation (p.i.), GH3:gusA expression was reduced in the vascular bundle between the inoculation site and the root tip (Fig. 1e; stage 1). After 10 h, increased GH3:gusA expression occurred immediately basipetally from the inoculation site in the cortex, and the reduced expression in the vascular bundle remained between the inoculation site and the root tip (Fig. 1f; stage 2). At 20 h p.i., GH3:gusA expression was high in a zone of about 2 mm to either side of the inoculation site in the cortex. At this time, expression in the vascular bundle began to reappear between the inoculation site and the root tip (Fig. 1g). However, the expression just basipetal from the root tip was in most cases lower than in control roots. At 30 h p.i., the root cortical and vascular cells showed enhanced GH3:gusA expression in a more extended zone around the inoculation bead than after 20 h (Fig. 1h; stage 4). Transverse sections at the site of inoculation confirmed high GH3:gusA expression in all cortical and vascular cells at the inoculation site (Fig. 2e). GH3:gusA expression in the vascular bundle occurred towards the root tip, but in 50% of the roots it remained visibly reduced compared to controls. After about 50 h, the first cell divisions were apparent in the inner cortex. At this stage, GH3:gusA expression was focused in the cells undergoing division, but was very low or not detectable in the surrounding non-dividing cortical cells (Fig. 1i and Fig. 2f; stage 5). By about 70 h, when the nodule primordium began to differentiate, no GH3:gusA expression was detected in the centre of the primordium, but expression remained at the base and periphery (Fig. 2g; stage 6). Examination of nodules 4–5 days after inoculation showed that GH3:gusA expression occurred in a bifurcated pattern (Fig. 2h; stage 7).

Non-nodulating Rhizobium strains do not alter GH3:gusA expression

To determine if the change in GH3:gusA expression was specifically induced by clover-nodulating bacteria, we spot-inoculated non-nodulating mutant and non-host bacteria onto the transgenic plants (Table 1). The inoculation of two non-nodulating ANU843 derivatives that are unable to synthesize LCOs, ANU277 (nodC::Tn5) and ANU845 (pSym-cured), did not induce any changes in GH3:gusA expression (Table 2). At 5 h p.i., the non-host R. meliloti strain Rm1021 initially induced similar changes to the pattern of GH3:gusA expression as that seen with ANU843; however, no further changes in GH3:gusA expression were seen at later time-points. The more distantly related bacterial species, Bradyrhizobium japonicum USDA110 and Agrobacterium tumefaciens C58, failed to induce any detectable changes in GH3:gusA expression (Table 2).

Table 1.  . Bacterial strains used for spot-inoculation Thumbnail image of
Table 2.  . Effects of bacterial strains and compounds on GH3:gusA expression patterns resulting from spot-inoculation Thumbnail image of

LCOs and O-acetylated chitin oligosaccharides cause a similar local downregulation of GH3:gusA expression as ANU843 and NPA

The microtargeting or spot-inoculation of control solutions (water or solvents) did not alter the expression of GH3:gusA (Fig. 3a). As positive controls, we examined the effect of indole-3-acetic acid (IAA), α-naphthaleneacetic acid (NAA), NPA and the ENOD40–1 peptide on GH3:gusA expression. The microtargeting of IAA and NAA increased GH3:gusA expression at and acropetal from the application site within 24 h of incubation (Fig. 3b and Table 3), whereas the spot-inoculation of IAA or NAA was ineffective (Table 2). Microtargeted or spot-inoculated NPA (Fig. 3c) induced strong downregulation of GH3:gusA expression acropetal from the inoculation site from 5 h onwards (Tables 2 and 3). Microtargeted MsENOD40–1 peptide (at 10–6, 10–7, 10–9 or 10–10m) did not induce any changes to GH3:gusA expression after 24 h (Fig. 3d, Table 3).


Figure 3 Alteration of GH3:gusA expression by microtargeted compounds after 24 h incubation. The target site is marked with red ink and indicated by an arrow. Roots were microtargeted with: (a) sterile water (left) (for comparison, an untreated root is shown on the right); (b) the auxin IAA (10–6m); (c) the synthetic auxin transport inhibitor, NPA (10–6m); (d) the MsENOD40–1 peptide (10–10m); (e) purified LCOs of strain ANU843 (10–8m) (these roots were incubated in a higher substrate concentration which resulted in a higher than normal expression, this was done to enhance the effect of local downregulation of GH3:gusA); (f) an O-acetylated chitin pentamer (10–6m) (these roots were incubated in a higher substrate concentration which resulted in a higher than normal expression, this was done to enhance the effect of local downregulation of GH3. :gusA); (g) an unsubstituted chitin pentamer (10–6m); (h) the flavonol quercetin (50 μm). Bars = 1 mm.

Download figure to PowerPoint

Table 3.  . Effects of compounds on GH3:gusA expression patterns, assayed 24 h after microtargeting Thumbnail image of

 Given that ANU843 induces changes in GH3:gusA expression during nodule initiation (Fig. 1e–i), we expected that LCOs inoculation would cause similar effects. Consequently, purified LCOs of strain ANU843 were either spot-inoculated onto or microtargeted into roots at 10–8m. The LCOs induced changes in GH3:gusA expression that were qualitatively and spatially similar to the changes in expression induced by ANU843 over the first 30 h p.i. (Fig. 1e–h and Fig. 3e), but had a slightly delayed onset (up to 5 h;Table 2). The local downregulation of GH3:gusA by LCOs was more enhanced in microtargeted roots compared to spot-inoculated roots, but in both cases resembled the effects observed with NPA. In contrast to results obtained with strain ANU843 inoculation, no further effect of LCOs application on GH3:gusA expression was apparent from 50 h p.i. onwards (Table 2). The microtargeting of LCOs into the mature root (several centimetres from the root tip) did not alter GH3:gusA expression (data not shown).

O-acetylated chitin oligosaccharides were microtargeted into roots because they form the LCOs backbone. Tetra-, penta- or hexamers of the O-acetylated chitin oligosaccharides induced a similar local downregulation of GH3:gusA expression as LCOs and NPA (Fig. 3f and Table 3). The pentamer had the strongest effect (Table 3). Microtargeted non-O-acetylated chitin oligosaccharides including either di-, tetra-, penta- or hexamers failed to induce alterations to GH3:gusA expression (Fig. 3g and Table 3).

Flavonoid aglycones cause similar changes to GH3:gusA expression as NPA

We tested the effect of flavonoids on GH3:gusA expression because certain flavonoids have been reported to act as endogenous auxin transport regulators (Jacobs and Rubery 1988). We observed that the structure of the microtargeted flavonoids affected the ability to influence GH3:gusA expression. The flavonols, quercetin (Fig. 3h), fisetin and kaempferol, the flavone, apigenin, and the flavanone, naringenin, caused a similar local and acropetal downregulation of GH3:gusA expression as LCOs, O-acetylated chitin fragments and NPA after 24 h (Table 3). Spot-inoculation of quercetin and apigenin had a similar effect as microtargeting (Table 2). In contrast to the aglycones, the microtargeting of isoflavonoid, genistein, and the flavonoid glucosides, quercetin-3-glucoside, kaempferol-3-glucoside, apigenin-7-rhamnoglucoside and naringenin-7-rhamnoglucoside, failed to alter GH3:gusA expression after 24 h (Table 3).


  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

How reliable is GH3:gusA expression as a marker for the activity of auxin?

A major limitation of plant biology is the inability to accurately assess the levels of active phytohormones in tissues and individual cells. We have used an indirect method to assess the auxin responsiveness of clover root tissues. We used the auxin responsive GH3 promoter that was originally isolated from soybean hypocotyls in conditions promoting auxin-stimulated cell elongation (Hagen et al. 1984). However, several results suggest that the transgenic white clover plants containing GH3:gusA are reliable indicators of changes in auxin activity during development. First, the expression patterns obtained for lateral root formation (Fig. 2) and gravi-stimulated roots (Larkin et al. 1996) are consistent with the physiological role of auxin in these processes, both temporally and spatially (Pelosi et al. 1995;Sussex et al. 1995). Furthermore, the expression pattern indicates that GH3:gusA expression is not restricted to cells during either auxin-stimulated division or elongation, but that every cell in the root is able to express GH3:gusA (Fig. 2e), as reported previously by Guilfoyle et al. (1993). Second, in vitro studies show the selectivity of the GH3 promoter for auxin and not other phytohormone classes (Larkin et al. 1996). In addition, the auxin inducibility increases linearly with rising auxin concentrations over a range from 10–6 to 10–3m (Guilfoyle et al. 1993). Third, the histochemical localization of the indigo dye product coincides with the immunolocalization of β-glucuronidase (Fig. 1a,b). Fourth, in situ hybridization of GH3 transcripts in soybean roots shows the vascular bundle, especially around the xylem poles, as the main site of GH3 mRNA accumulation (Guilfoyle et al. 1993), similar to GH3:gusA expression patterns described in this study. Furthermore, as would be expected, GH3:gusA was upregulated by local application of auxins (Fig. 3b) and acropetally downregulated by the auxin transport inhibitor, NPA (Fig. 3c), and flavonoid aglycones, but not flavonoid glycosides (Jacobs and Rubery 1988). Therefore, interpretations about the effects of bacteria and signal molecules used in this study were based on comparisons with the effect of auxins and established auxin transport inhibitors. Notwithstanding these findings, it is possible that potential artifacts may occur due to indirect detection of the GUS enzyme (Guivarc’h et al. 1996), the nature of the GH3 regulatory regions (Taylor 1997) or the differential perception of auxin by cells via separate signalling pathways (Guilfoyle et al. 1993). Therefore, future studies will attempt to verify, by different methods, that auxin transport inhibition occurs early during root nodule formation.

Auxin transport inhibition precedes the earliest stages of root nodule formation

Rhizobium inoculation leads to temporal and spatial changes in GH3:gusA expression in white clover roots. Using the inoculation site as a reference point, two main patterns of GH3:gusA expression were recognized before initiation of cell divisions (stages 1–4). A rapidly induced (stage 1), local and transient (until stage 2) downregulation of GH3:gusA expression occurred in the vascular bundle and cortex (Fig. 1e,f), followed by a basipetal upregulation in the cortex and vascular bundle (stage 2–4;Fig. 1f–h). Both acropetal downregulation and basipetal upregulation of GH3:gusA was mimicked by the auxin transport inhibitor NPA, flavonoid aglycones and LCOs (Tables 2 and 3 and Fig. 3c,e,h).

 We conclude from a comparison of the effects of auxins and NPA treatments with that of Rhizobium inoculation and LCOs addition, that an inhibition of the acropetal auxin flow leads to the downregulation of GH3:gusA expression between the inoculation site and the tip. Concomitantly, the basipetal upregulation of GH3:gusA is most probably due to an accumulation of shoot-derived auxin that is stopped at the inoculation site. By 20–30 h p.i., the auxin transport inhibition presumably weakens, because GH3:gusA gradually reappears acropetally from the inoculation site, although at a slower rate in the cortex than in the vascular bundle (Fig. 1g,h). However, even at 30 h p.i., GH3:gusA expression is still reduced behind the root tip, compared to untreated controls.

 In contrast to ANU843, non-nodulating ANU843 derivatives (which do not make LCOs) and non-host rhizobia (which make incorrect LCOs) were unable to induce sustained auxin transport inhibition (Table 2). These results suggest that local auxin transport inhibition is necessary for nodule initiation. The exceptional result was the change in GH3:gusA expression that occurred 5 h after the inoculation with R. meliloti, which did not occur at later time-points. Because the inoculation of R. meliloti strain 1021 led to root hair distortions at the inoculation site but not to root nodules or infection threads, this finding may suggest that the R. meliloti LCOs are recognized by clover roots but that all of the changes needed to induce nodule initiation do not occur. Indeed, recent studies have shown that non-legume roots show some responsiveness to LCOs (Spaink 1996).

LCOs, O-acetylated chitin fragments and flavonoid aglycones induce auxin transport inhibition

LCOs, the necessary signal molecules of rhizobia, are the most likely mediators of the effect of ANU843 on auxin transport, because LCOs caused similar changes in GH3:gusA expression as application of ANU843 and of NPA (Tables 2 and 3 and Fig. 3e). This would be expected if auxin transport inhibition was a prerequisite for nodulation. Interestingly, O-acetylated chitin oligosaccharides induced similar changes in GH3:gusA expression as LCOs (Fig. 3f), indicating that the chitin backbone of LCOs is the biologically active moiety. Recently, O-acetylated chitin fragments were shown to induce cortical cell divisions in roots of Vicia sativa following microtargeting (Schlaman et al. 1997). Our results indicate that LCOs are especially active in the susceptible zone near the root tip, because the microtargeting of LCOs into the mature part of the root did not induce changes in GH3:gusA expression. Bhuvaneswari et al. (1981) showed that the area behind the root tip was particularly sensitive to Rhizobium inoculation.

 Flavonoids are potential candidates for endogenous inhibitors of auxin transport in the early stages of nodule initiation (Hirsch et al. 1989;Hirsch 1992). Our results support this hypothesis, because introduction of flavonoids into the roots had a similar effect upon GH3:gusA expression as LCOs, O-acetylated chitin fragments and NPA (Fig. 3h). Flavonols, flavanones and flavones were most active, whereas their respective glycosides and also an isoflavonoid were inactive (Table 3), which indicates a specificity for certain flavonoid aglycones as auxin transport inhibitors. Similar results were obtained by Jacobs, Rubery (1988), who used the same flavonoids in their study. Also consistent with the biological activity of flavonoids is the specific induction of flavonoid genes and end products by compatible rhizobia on their respective hosts (clover;Lawson et al. 1994, 1996;Vicia, Recourt et al. 1992). We have obtained additional data to support these findings (U. Mathesius et al. unpublished data).

Changes in GH3:gusA expression during nodule differentiation are similar to changes during lateral root differentiation

Only the clover nodulating strain ANU843 (Fig. 2e−h) was able to induce stage 1–6 changes in GH3:gusA expression, whereas R.l. bv. trifolii LCOs were only able to induce the stage 1–4 changes (Table 2). The activity of plant chitinases (Staehelin et al. 1994) could explain why LCOs alone were unable to induce nodule initiation. In contrast, compatible bacteria continually produce LCOs during the infection process. Between stage 4 and 5, a reduction in GH3:gusA expression occurred in all cortex cells except a specific subset of inner cortical cells (Fig. 2f) that continued to divide to form the nodule primordium. The specific retention of GH3:gusA expression in these dividing inner cortical cells may infer that an additional signal is required to trigger these particular cells to divide. One possible signal, the ENOD40–1 peptide, is reported to change the response of cells to auxin (van de Sande et al. 1996). However, our results did not support any role for the ENOD40–1 peptide in perturbing the auxin balance when added exogenously (Fig. 3d).

 Because high levels of GH3:gusA expression were observed before initiation of cell division (Fig. 2e) and in the progenitor cells of a nodule primordium (Fig. 2f), a period of high auxin exposure and responsiveness is likely to be necessary in these cells prior to and during the first divisions (stages 4 and 5). Since it has also been found that cytokinin application can induce nodule formation (Bauer et al. 1996;Cooper and Long 1994), it is unlikely that the auxin levels alone determine the conditions for nodule initiation, but that the ratio of cytokinin to auxin levels is important for nodule initiation. Similar to nodule progenitor cells, high GH3:gusA expression was also detected during lateral root initiation in pericycle cells prior to and during division (Figs 1d and 2a). The difference between progenitor cells of nodule and lateral root primordia is that pericycle cells were always situated in a zone of high GH3:gusA expression and did not undergo a transient phase of GH3:gusA downregulation (i.e. no stages 1–2). High auxin requirements in cells prior to division could reflect the role of auxin in the accumulation of p34cdc2-like proteins needed to activate mitosis (John et al. 1993;Zhang et al. 1996).

 After the initiation of nodule and lateral root primordia, we found further similarities in GH3:gusA expression patterns at parallel developmental stages. In both organs, expression was high in dividing cells of the early primordium (Fig. 2b,f) but reduced during later stages of development and differentiation (Fig. 2c,d,g,h). These observations are supported by the results of Pelosi et al. (1995) and Sussex et al. (1995), which show that lateral root morphogenesis is a two-step process (requiring high auxin concentration during the initiation of the primordium and subsequently lower auxin concentrations during primordium differentiation). After the initiation of the lateral root primordium, a pulse of GH3:gusA expression was seen in the outer cortex overlying the primordium. This resolves the issue raised in Larkin et al. (1996), where it was unclear whether this characteristic pulse preceded or followed lateral root initiation. Because the GH3:gusA expression patterns in developing nodules are very similar to patterns during lateral root formation, our results suggest that nodule formation also requires high auxin levels for initiation of cell divisions (Fig. 2e) and establishment of the primordium (Fig. 2f), but lower levels after the establishment of the nodule primordium (Fig. 2g,h).

 Overall, our results show for the first time that rhizobia cause a localized, temporary and early inhibition of auxin transport, which subsequently leads to an accumulation of auxin at the site of nodule initiation. Although the later stages of nodule and lateral root formation are similar in their apparent requirements for auxin, the changes in GH3:gusA expression preceding nodule initiation are unique. Both lateral root and nodule formation appear to be a two-step process in their auxin requirement. Possible endogenous signal molecules that could mediate the effect of LCOs on auxin transport are certain flavonoid aglycones. Current experiments also indicate that localized flavonoid induction occurs inside the root following Rhizobium inoculation (U. Mathesius et al. unpublished data). Our results also show that the O-acetylated chitin moiety of the LCO molecule is biologically active.

Experimental procedures

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Bacterial strains and test compounds

Table 1 lists the bacterial strains used. Test compounds used for spot-inoculation or microtargeting were NPA, NAA and IAA (Sigma Chemicals Co., St Louis, MO), dissolved in 0.01 or 1% methanol; and kaempferol, apigenin, apigenin-7-rhamnoglucoside, genistein (Roth, Karlsruhe, Germany), quercetin, naringenin, naringenin-7-rhamnoglucoside (Sigma), quercetin-3-glucoside, kaempferol-3-glucoside (Indofine Chemicals, France), dissolved in 5% methanol. The MsENOD40–1 peptide was provided by Henk Franssen and dissolved in 5% DMSO. Chitin fragments composed of five or six glucosamine residues were obtained from Seika-Gaku (Tokyo, Japan) and dissolved in Milli-Q purified water. Chitin fragments were O-acetylated at the C6 position of the non-reducing sugar using the NodL enzymatic assay, and purified using HPLC as described previously (Bloemberg et al. 1994). LCOs were dissolved in 1% DMSO or 5 mm cyclodextrin (Sigma).

Transgenic plants

The transgenic white clover (Trifolium repens cv. Haifa) plants used have been described previously (Larkin et al. 1996). Five first generation transgenic lines were used for analysis. Mature first generation transgenic plants were grown in soil under natural light conditions at 26°C, 60% average humidity during the day, and 19°C, 80% average humidity during the night, and were maintained by regrowth from shoot cuttings.

Generation of rooted leaves

Axenically grown rooted leaves were generated from mature plants (Rolfe and McIver 1996) using Jensen's medium (Jensen 1942). The roots were kept dark by covering the lower half of the plates with brown paper or specifically designed black plastic boxes. Growth conditions for the rooted leaves were, 16 h day (24–25°C), 8 h night (19°C) with a mixture of fluorescent and incandescent light with an intensity of ∼ 140 μmol m–2 s–1 and 80% average humidity. Under these conditions, roots formed visible nodules after 4–5 days p.i.

Spot-inoculation of transgenic plants

Two transgenic lines were primarily used for the experiments. Of each line, 5–20 rooted leaves were used for each treatment and time-point. Equal numbers of roots were inoculated with either a test or a (negative) control solution consisting of the solvent only. In the case of bacterial inoculations, Bergensens Modified Medium (BMM) (Rolfe et al. 1980) was used as a negative control. Positive controls consisted of roots inoculated with R. l. bv. trifolii. One rooted leaf each with one to four roots exceeding 1 cm (8–10 days growth) was transferred to a fresh, dry Jensen's plate and grown overnight. The fresh dry plate was needed to avoid spreading of the inoculum due to excessive moisture. The spot-inoculation technique of Bhuvaneswari et al. (1981) was used. One root per rooted leaf was spot-inoculated with ∼ 20 nl of the dissolved compound or solvent. For bacterial inoculations, bacteria were grown in liquid BMM at 28°C over night, to OD (600 nm) of 0.2–0.4 before inoculation. Bacteria, 5000–10 000, (as tested by plate counts) were applied thus avoiding any effects of over-inoculation.

Microtargeting of transgenic plants

Microtargeting was carried out with a modified microtargeting microprojectile accelerator (Gisel et al. 1996;Sautter et al. 1991). Each test compound was dissolved in solvent at the final concentration and mixed with an equal volume of gold particles (average size of 1 μm diameter), suspended in sterile water. To introduce the aerosol, a 60 nl suspension load with a restriction size of 140 μm was used at 50 bar nitrogen. Rooted leaves with 1 cm long roots were used. The leaf was excised 5 mm above the emerged roots and the roots mounted on Jensen's plates containing 2% agarose with 2% maltose as an osmoticum, and incubated for 2 h prior to treatment. One root per root system was microtargeted in the same area used for spot-inoculation and that spot was marked with sterile red ink. Control solutions were microtargeted into an equal number of roots as the test compounds. After microprojectile bombardment, the roots were incubated a further 24 h at 25°C in a horizontal position and subsequently used for histochemical GUS localization.

Isolation of LCOs

LCOs were isolated from Rhizobium leguminosarum bv. trifolii strain ANU843 either following the method of Spaink et al. (1995) or as a modification of this technique by L. F. Roddam, J. W. Redmond and M. A. Djordjevic (manuscript in preparation).

GUS assays

Histochemical GUS assays were carried out as described in Larkin et al. (1996).


Whole, histochemically stained roots were examined using a Leica or a Nikon SMZ-10 stereomicroscope. For sectioning, fixed and histochemically stained material was embedded in 3% agarose and sectioned (80 μm thickness) on a vibratome (Lancer series 1000). Alternatively, roots were fixed overnight at 4°C in 2.5% glutaraldehyde, 2% paraformaldehyde in 0.1 m cacodylate buffer (pH 7.2), dehydrated in increasing concentrations of ethanol and infiltrated with historesin following the manufacturer's protocol (Historesin Embedding Kit, Jung). Sections of 10 μm were cut on a Reichert Jung Ultra Microtome and analysed and photographed using a Nikon Optiphot light microscope fitted with a Nikon FX 35 camera. Kodak Ektachrome 100 ASA or EPY 64 T colour slide film was used.


Histochemically stained roots were fixed in 4% paraformaldehyde in phosphate buffer at 4°C over night and embedded in historesin as described above. Root sections were washed with BSA buffer (containing 20 mm Tris–HCl, 0.1% BSA, 0.02% fish skin gelatine) and incubated with a polyclonal rabbit anti-β-glucuronidase antiserum (Clontech Laboratories) overnight in BSA buffer at 4°C. Sections were rinsed with BSA buffer and incubated for 2 h with a secondary pig anti-rabbit–FITC labelled antibody (Dako, Denmark) and viewed and photographed using a Leitz Diaplan fluorescent microscope. Kodak 400 ASA film was used for fluorescent photographs.


  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

U. Mathesius is a recipient of an Overseas Postgraduate Research Award from the Australian Government. H. R. M. Schlaman was supported by a Pionier Grant from the Netherlands Organization for Scientific Research (NWO) awarded to H. P. Spaink. Professor M. E. McCully (Carleton University, Ottawa, Canada) is thanked for many discussions and suggestions, and Elena Gärtner (ANU) for technical support. Thanks to Daphne Meijer and Pieter Admiraal (microtargeting experiments) and Gerda Lamers (Leiden University) for immunohistochemistry. We thank Henk Franssen (Wageningen Agricultural University) for the MsENOD40–1 peptide.


  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  • 7
    Bauer, P., Ratet, P., Crespi, M.D., Schultze, M., Kondorosi, A. 1996 Nod-factors and cytokinins induce similar cortical cell divisions, amyloplast deposition and MsENOD12A expression patterns in alfalfa roots. Plant J., 10, 91 105.
  • 20
    Bhuvaneswari, T.V., Bhagwat, A.A., Bauer, W.D. 1981 Transient susceptibility of root cells in four common legumes to nodulation by rhizobia. Plant Physiol., 68, 1144 1149.
  • 31
    Bloemberg, G.V., Thomas-Oates, J.E., Lugtenberg, B.J.J., Spaink, H.P. 1994 Nodulation protein NodL of Rhizobium leguminosarum O-acetylates lipo-oligosaccharides, chitin fragments and N-acetylglucosamine. Mol. Microbiol., 11, 793 804.
  • 8
    Cooper, J.B. & Long, S.R. 1994 Morphogenetic rescue of Rhizobium meliloti nodulation mutants by trans-zeatin secretion. Plant Cell, 6, 215 225.
  • 1
    Dénarié, J., Debellé, F., Promé, J.C. 1996 Rhizobium lipo-chito-oligosaccharide nodulation factors: signaling molecules mediating recognition and morphogenesis. Annu. Rev. Biochem., 65, 503 535.
  • 37
    Djordjevic, M.A., Schofield, P.R., Rolfe, B.G. 1985 Tn5 mutagenesis of Rhizobium trifolii host-specific nodulation genes result in mutants with altered host-range ability. Mol. Gen. Genet., 200, 463 471.
  • 35
    Gisel, A., Iglesias, V.A., Sautter, C. 1996 Ballistic microtargeting of DNA and biological active substances to plant tissues. Plant Tiss. Cult. Biotechnol., 2, 33 41.
  • 24
    Guilfoyle, T.J., Hagen, G., Li, Y., Ulmasov, T., Liu, Z., Strabala, T., Gee, M. 1993 Auxin-regulated transcription. Aust. J. Plant Physiol., 20, 489 502.
  • 25
    Guivarc’h, A., Caissard, J.C., Azmi, A., Elmayan, T., Chriqui, D., Tepfer, M. 1996 In situ detection of expression of the GUS reporter gene in transgenic plants: ten years of blue genes. Transgenic Res., 5, 281 288.
  • 15
    Hagen, G., Kleinschmidt, A., Guilfoyle, T. 1984 Auxin-regulated gene expression in intact soybean hypocotyl and excised hypocotyl sections. Planta, 162, 147 153.
  • 4
    Hirsch, A.M. 1992 Developmental biology of legume nodulation. New Phytol., 122, 211 237.
  • 9
    Hirsch, A.M., Bhuvaneswari, T.V., Torrey, J.G., Bisseling, T. 1989 Early nodulin genes are induced in alfalfa root outgrowths elicited by auxin transport inhibitors. Proc. Natl. Acad. Sci. USA, 86, 1244 1248.
  • 10
    Jacobs, M. & Rubery, P.H. 1988 Naturally occurring auxin transport regulators. Science, 241, 346 349.
  • 17
    Jefferson, R.A., Kavanagh, T.A., Bevan, M.W. 1987 GUS fusions: β-glucuronidase as a sensitive and versatile gene fusion marker in higher plants. EMBO J., 6, 3901 3907.
  • 33
    Jensen, H.L. 1942 Nitrogen fixation in leguminous plants. I. General characters of root nodule bacteria isolated from species of Medicago and Trifolium in Australia. Proc. Linn. Soc. NSW, 67, 98 108.
  • 12
    John, P.C.L., Zhang, K., Dong, C., Diederich, L., Wightman, F. 1993 p34 cdc2 related proteins in control of cell cycle progression, the switch between division and differentiation in tissue development, and stimulation of division by auxin and cytokinin. Aust. J. Plant Physiol., 20, 503 526.
  • 14
    Larkin, P.J., Gibson, J.M., Mathesius, U., Weinman, J.J., Gärtner, E., Hall, E., Tanner, G.J., Rolfe, B.G., Djordjevic, M.A. 1996 Transgenic white clover. Studies with the auxin responsive promoter, GH3, in root gravitropism and lateral root development. Transgenic Res., 5, 325 335.
  • 27
    Lawson, C.G.R., Djordjevic, M.A., Weinman, J.J., Rolfe, B.G. 1994 Rhizobium inoculation and physical wounding result in the rapid induction of the same chalcone synthase copy in Trifolium subterraneum. Molec. Plant Microbe Interact., 7, 498 507.
  • 28
    Lawson, C.G.R., Rolfe, B.G., Djordjevic, M.A. 1996 Rhizobium inoculation induces condition dependent changes in the flavonoid composition of root exudates from Trifolium subterraneum. Aust. J. Plant Physiol., 23, 93 101.
  • 6
    Libbenga, K.R., Van Iren, F., Bogers, R.J., Schraag-Lamers, M.F. 1973 The role of hormones and gradients in the initiation of cortex proliferation and nodule formation in Pisum sativum L. Planta, 114, 29 39.
  • 38
    Meade, H.M., Long, S.R., Ruvkun, G.B., Brown, S.E., Ausubel, F.M. 1982 Physical and genetic characterisation of symbiotic and auxotrophic mutants of Rhizobium meliloti induced by transposon Tn5 mutagenesis. J. Bacteriol., 149, 114 122.
  • 16
    Napier, R.M. & Venis, M.A. 1995 Auxin action and auxin binding proteins. New Phytol., 129, 167 201.
  • 39
    Nieuwkoop, A.J., Banfalvi, Z., Deshanme, N., Gerhold, G., Schell, M.G., Sirotkin, K.M., Stacey, G. 1987 A locus encoding host range is linked to the common nodulation genes. J. Bacteriol., 169, 2631 2638.
  • 23
    Pelosi, A., Lee, M.C.S., Chandler, S.F., Hamill, J.D. 1995 Hormonal control of root primordia differentiation and root formation in cultured explants of Eucalyptus globulus seedlings. Aust. J. Plant Physiol., 22, 409 415.
  • 29
    Recourt, K., Van Tunen, A.J., Mur, L.A., Van Brussel, A.A.N., Lugtenberg, B.J.J., Kijne, J.W. 1992 Activation of flavonoid biosynthesis in roots of Vicia sativa subsp. nigra plants by inoculation with Rhizobium leguminosarum biovar viciae. Plant Mol. Biol., 19, 211 220.
  • 34
    Rolfe, B.G., Gresshoff, P.M., Shine, J. 1980 Rapid screening for symbiotic mutants of Rhizobium and white clover. Plant Sci. Lett., 19, 277 284.
  • 32
    Rolfe, B.G. & McIver, J. 1996 Single-leaf plantlet bioassays for the study of root morphogenesis and Rhizobium nodulation. Aust. J. Plant Physiol., 23, 271 283.
  • 11
    Van De Sande, K., Pawlowski, K., Czaja, I., Wieneke, U., Schell, J., Schmidt, J., Walden, R., Franssen, H., Bisseling, T. 1996 Modification of phytohormone response by a peptide encoded by ENOD40 of legumes and a non legume. Science, 273, 370 373.
  • 21
    Sautter, C., Waldner, H., Neuhaus-Url, G., Galli, A., Neuhaus, G., Potrykus, I. 1991 Micro-targeting: high efficiency gene transfer using a novel approach for the acceleration of micro-projectiles. Bio/Technol., 9, 1080 1085.
  • 22
    Schlaman, H.R.M., Gisel, A.A., Quaedvlieg, N.E.M., Bloemberg, G.V., Lugtenberg, B.J.J., Kijne, J.W., Potrykus, I., Spaink, H.P., Sautter, C. 1997 Chitin oligosaccharides can induce cortical cell division in roots of Vicia sativa when delivered by ballistic microtargeting. Development, 124, 4887 4895.
  • 40
    Schofield, P.R., Djordjevic, M.A., Rolfe, B.G., Shine, J., Watson, J.M. 1983 A molecular linkage map of nitrogenase and nodulation genes in Rhizobium trifolii. Mol. Gen. Genet., 192, 459 465.
  • 2
    Schultze, M., Kondorosi, E., Ratet, P., Buiré, M., Kondorosi, A. 1994 Cell and molecular biology of Rhizobium–plant interactions. Internat. Rev. Cytol., 156, 1 75.
  • 3
    Spaink, H.P. 1996 Regulation of plant morphogenesis by lipo-chitin oligosaccharides. Crit. Rev. Plant Sci., 15, 559 582.
  • 36
    Spaink, H.P., Bloemberg, G.V., Van Brussel, A.A.N., Lugtenberg, B.J.J., Van Der Drift, K.M.G.M., Haverkamp, J., Thomas-Oates, J.E. 1995 Host specificity of Rhizobium leguminosarum is determined by the hydrophobicity of highly unsaturated fatty acyl moieties of the nodulation factors. Molec. Plant Microbe Interact., 8, 155 164.
  • 30
    Staehelin, C., Schultze, M., Kondorosi, E., Mellor, R.B., Boller, T., Kondorosi, A. 1994 Structural modifications in Rhizobium meliloti Nod factors influence their stability against hydrolysis by root chitinases. Plant J., 5, 319 330.
  • 18
    Sussex, I.M., Godoy, J.A., Kerk, N.M., Laskowski, M.J., Nusbaum, H.C., Welsch, J.A., Williams, M.E. 1995 Cellular and molecular events in a newly organising lateral root meristem. Phil. Transac. Royal Soc. Lon., 359, 39 43.
  • 26
    Taylor, C.B. 1997 Promoter fusion analysis: an insufficient measure of gene expression. Plant Cell, 9, 273 275.
  • 19
    Wightman, F., Schneider, E.A., Thimann, K.V. 1980 Hormonal factors controlling the initiation and development of lateral roots. Physiol. Plant., 49, 304 314.
  • 5
    Yang, W.C., De Blank, C., Meskiene, I., Hirt, H., Bakker, J., Van Kammen, A., Franssen, H., Bisseling, T. 1994 Rhizobium Nod-factors reactivate the cell cycle during infection and nodule primordium formation, but the cell cycle is only completed in the primordium formation. Plant Cell, 6, 1415 1426.
  • 13
    Zhang, K., Letham, D.S., John, P.C.L. 1996 Cytokinin controls the cell cycle at mitosis by stimulating the tyrosine dephosphorylation and activation of p34cdc2-like H1 histone kinase. Planta, 200, 2 12.
  1. e-mail